Wednesday, January 11, 2012

Graphics & Physics Update

Just a quick update to show some progress. Here's a video:



The rundown:
  • Added graphics for hand (temporary)
  • Refined hand physics (it now follows a pivot which moves with the controller, much better performance
  • Moved around some classes and physics engine classifications for objects, the code's neater
  • Changed how the hand receives information from the controller, works better
  • Prepped the code for adding a new type of controller
  • Validated bumpmaps are working in material imports


Which brings us to our last point, the next objective: Add a keyboard/mouse controller. Still working on how exactly that'll be, but I have some ideas. Hopefully in addition to videos next time I'll have a downloadable demo available on here for it. Also want to work on a level loader in hopes that we'll have a working level soon.

Monday, December 26, 2011

PhysHand Preview

Just finished some incredibly entertaining work on a physics-manipulation hand, and it's a lot of fun to play with. Here's a video:



More important than being fun, with very little time spent with it you get pretty used to judging depth in the game, which is awesome. After working on it for an hour I wasn't having any trouble picking up a specific object, despite the lack of visual for where the hand is. I'll probably add a visual in eventually, but that I can use it without one is very encouraging.

Thursday, December 8, 2011

Decoupling

I just finished reading an article in this month's Game Developer magazine about decoupling systems.  In it, Niklas Frykholm talks about the methods and reasons to seperate systems in a game engine.  He has a very cool bit about duck typing in C++ with an array-of-structs design, which is really cool and I'm going to have to try.  He ends the article with a quip about how he wants APIs to be: C or barebones C++ interfaces that don't require inheritance.

That's probably the reason I like the sixense sdk so much; it exposes a small, simple interface to my code, and doesn't try to handle anything that isn't reading data from the controllers.  It's very easy and uninvasive to work with.

Speaking of, the project is still moving, albiet slowly.  Leo and I have both been very busy, but despite that I've gotten more done on the program I was demonstrating earlier, and got a second prototype going.  Right now I'm working on a "physics hand" to allow manipulation of objects by the player.  It's going to use the mass the object has in the Leadwerks physics system to determine if you can lift something, just drag it around, or if you can't budge it at all.  This should allow fairly intuitive integration into levels, as the mass should already be set for objects that have physics anyway.  Once it's done, I'll post another demo with it.

Graphics are importing just fine, but unfortunately we don't have many so far, so the next video might be white cubes again.  Maybe I'll add different color cubes.  And I promise I'll use CamStudio this time.

Thursday, October 27, 2011

Playing with Some New Guns

Leo here.

So I've decided that sleep wasn't going to be something that I would do anymore, and spend my nights researching and playing around with some modeling and level design software to get things rolling and tested.

The first level design software I was suggested was 3D World Studio.

I've done a lot of research on this, and there was something they said that caught my eye:



"3D World Studio offers a feature set just as powerful as UnrealEd, yet is friendly enough for beginners to use."


A level editor I've never heard of that claims to rival UnrealEd is almost /unreal/ to me (get it? Shut up).

So I delve into this monster and see what she's made of:

First off, what I like:
Most of the editor functions like UDK would, but the feature I could applaud is the carve feature. It takes any shape you create, and obliterates everything that intersects with it. Of course, UDK did this too, but you would have to create a specific subtraction BSP and sometimes it would even cut the BSP weird because of irregular subtraction BSP. In 3DSW, it's basically just: What do you want your obliterator to look like? Where would you like it to be? OBLITERATE!

So far, it hasn't caused me much trouble, and the calculations seem to be dead on. Then again, I've only been making the most obscure shapes and using the most uninhabitable structures to carve it out. Needless to say, I get interesting results.

What I don't like:
It's extremely difficult to get my own objects into the editor, and to add insult to injury, it never scales the object correctly. When I say this, I don't mean "Oh, I made a mistake and it didn't scale correctly". I mean I exported one model, then exported the same model at x10 the size, but it always imports into the editor at the same size.

Another thing that is disappointing about 3DSW is the texturing. It understands which texture I want to use for the object, which is cool, but it only ever registers the diffuse map even though I've added so much detail in an effort to make it look sleek. All of my effort gone down the drain, and 3DSW throws a papercraft version of my model back in my face.





Flat Textures make me sad face :(


From my understanding specular, and normals are to be programmed in when I pass the map onto Bill, but I guess it's my lack of understanding of Leadwerks that leads me to believe that sounds like a pain in the ass.



Of course, the Leadwerks Sandbox is also an option, because it supports multiple textures on an object, but the documentation says I have to do some programming... I hardly think I want to spend the time doing that. Not only that, but it there isn't a way to create BSPs.

Unfortunately, Documentation for 3DWS is very much lacking aside from the user manual, which isn't very much help.

Looks like I'm gonna have some more studying to do. I thought I was done with school!

LATER! (That's my sign off now.)

Wednesday, October 19, 2011

Project Concept and Background

Two years ago, after just graduating college, I attended the GDC 2010 with a few of my friends, and while there I saw the Sixense TrueMotion system being demoed in Left 4 Dead 2.  It was awesome; full one-to-one control, six degrees of freedom, the whole bit.  I must've spent an hour talking to the people there about how it detects location with a weak magnetic field, how its SDK works and the like.

Well, I was very excited to see it release as the Razer Hydra, and even more excited to see the SDK be released with it.  And the SDK is beautiful.  So I got it, and shortly after placing the order considered what I'd do with it.

I'd used the Leadwerks Engine for my junior project in college, and was very, very happy with how it worked.  The engine looked good, the physics were solid, the lighting was completely dynamic.  Very cool.  Even better, the API was very nice, and it lacked all of the "frills" that, which nice when working with a team, I often felt just got in the way.

So, the idea was that I'd use those two things together to make something.  So, for starters, I'm making a series of Proofs of Concept of different ideas of how to use it.  This will be the place where I post updates for it.  I've been talking to Leo, and he's going to be providing some graphics.

And, for starters, a very low-quality video demonstrating some of what I have so far.




More to come.